On source code completion assistants and the need of a context-aware approach

Página do item simplificado
dc.contributor.authorARREBOLA, F. V.
dc.contributor.authorPlinio Thomaz Aquino Junior
dc.contributor.authorOrcidhttps://orcid.org/0000-0002-5100-7443
dc.date.accessioned2022-10-01T06:05:24Z
dc.date.available2022-10-01T06:05:24Z
dc.date.issued2017-07-09
dc.description.abstract© Springer International Publishing AG 2017.Source code completion assistance is a popular feature in modern IDEs. However, despite their popularity, there is little research about their key characteristics and limitations. There is also little research about the way software developers interact with code completion assistants, especially when considering the different techniques assistants use to populate the list of possible completions. This paper presents a study about the features of currently available code assistants and an experiment targeting professional Java developers familiar with the Eclipse platform that aims to collect and interpret usage data of two popular code completion assistants during the execution of three programming tasks. Results indicate that half the interactions with code assistants are either dismissed, interrupted or the completion proposals displayed have no direct contribution to the completion of the programming task. In that sense, we argue that code assistants still have a long road to pursue, since they seem to diminish the importance of the ultimate goals of the task at hand and also lack the ability of identifying and exploring the concepts of context-aware computing theory. The results of this paper can drive future HCI research to the design of adaptive code completion assistants that are able to respond to end user behaviors and preferences.
dc.description.firstpage191
dc.description.lastpage201
dc.description.volume10274 LNCS
dc.identifier.citationARREBOLA, F. V.; AQUINO JUNIOR, A. F. V. On source code completion assistants and the need of a context-aware approach. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), v.10274 LNCS, p. 191-201, July, 2017.
dc.identifier.doi10.1007/978-3-319-58524-6_17
dc.identifier.issn1611-3349
dc.identifier.urihttps://repositorio.fei.edu.br/handle/FEI/4596
dc.relation.ispartofLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
dc.rightsAcesso Restrito
dc.subject.otherlanguageDecision support systems
dc.subject.otherlanguageEvaluating information
dc.subject.otherlanguageIntelligent systems
dc.subject.otherlanguageKnowledge management
dc.titleOn source code completion assistants and the need of a context-aware approach
dc.typeArtigo de evento
fei.scopus.citations3
fei.scopus.eid2-s2.0-85025130765
fei.scopus.subjectCode completions
fei.scopus.subjectContext-aware approaches
fei.scopus.subjectContext-aware computing
fei.scopus.subjectEvaluating information
fei.scopus.subjectJava developers
fei.scopus.subjectKey characteristics
fei.scopus.subjectProgramming tasks
fei.scopus.subjectSoftware developer
fei.scopus.updated2025-02-01
fei.scopus.urlhttps://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85025130765&origin=inward
Arquivos
Coleções
Artigos